ETD PDF

Clustered Autoencoder Imputation

Citation

Furman, Daniel. (2020-05). Clustered Autoencoder Imputation. Theses and Dissertations Collection, University of Idaho Library Digital Collections. https://www.lib.uidaho.edu/digital/etd/items/furman_idaho_0089n_11845.html

Title:
Clustered Autoencoder Imputation
Author:
Furman, Daniel
Date:
2020-05
Keywords:
autoencoder clustering data pre-processing imputation machine learning neural networks
Program:
Mathematics
Subject Category:
Applied mathematics; Mathematics; Computer science
Abstract:

Many datasets have missing entries. Since downstream tasks often require full datasets with little

noise, accurately imputing the missing data is quite valuable. Autoencoders have proven themselves as

effective data imputers. However, while they exploit high order dependencies between the columns of

a dataset, autoencoders typically treat each row independently. This produces two problems. First,

imputation accuracy is suboptimal because not all of the data is used effectively. Second, downstream

classification tasks suffer since rows belonging to different classes get treated the same. Presented in this

thesis is CLAIM (CLustered Autoencoder IMputation), an algorithm that adapts existing autoencoder

networks in a way that directly addresses these issues. CLAIM first separates rows into clusters based

on similarity. Then, in the encoder, it applies different, loosely connected, learned linear transformations

to each cluster. Results show that this method improves accuracy with typical autoencoder imputation

strategies on large enough datasets. Also presented is a CLAIM-specific iterative clustering algorithm,

which allows CLAIM to improve initial cluster assignments as needed.

Description:
masters, M.S., Mathematics -- University of Idaho - College of Graduate Studies, 2020-05
Major Professor:
Gao, Fuchang
Committee:
Nguyen, Linh; Krone, Stephen
Defense Date:
2020-05
Identifier:
Furman_idaho_0089N_11845
Type:
Text
Format Original:
PDF
Format:
application/pdf

Contact us about this record

Rights
Rights:
In Copyright - Educational Use Permitted. For more information, please contact University of Idaho Library Special Collections and Archives Department at libspec@uidaho.edu.
Standardized Rights:
http://rightsstatements.org/vocab/InC-EDU/1.0/